PubMed Cell/Gene-Relation Atlas (PuMA)

Currenlty, please cite the following paper:

Plain

PuMA: PubMed Gene-Celltype-Relation Atlas
Lucas Bickmann, Sarah Sandmann, Carolin Walter, Julian Varghese
bioRxiv 2024.02.09.579584; doi: https://doi.org/10.1101/2024.02.09.579584

Bibtex

@article {PuMA,
author = {Lucas Bickmann and Sarah Sandmann and Carolin Walter and Julian Varghese},
title = {PuMA: PubMed Gene-Celltype-Relation Atlas},
elocation-id = {2024.02.09.579584},
year = {2024},
doi = {10.1101/2024.02.09.579584},
publisher = {Cold Spring Harbor Laboratory},
URL = {https://www.biorxiv.org/content/early/2024/02/12/2024.02.09.579584},
eprint = {https://www.biorxiv.org/content/early/2024/02/12/2024.02.09.579584.full.pdf},
journal = {bioRxiv}
}

How to:

1. Select the appropriate navigation element (Genes, Cells, Graphs).
2. Enter the search term
3. Select if you want an exact match and a filter for the minimal score/number of nodes (for Graphs)
4. Press Search

You can filter, sort and export the results with the appropiate described element. The graph view allows interactions.

Score information:

There exists two metrics: #PubMedIDs and #Occurences.
#PubMedIDs is the number where at least one occurence for the specific CUI is made.
#Occurence is the number of occurence for all PubMedIDs.

PuMA uses the following scoring scheme:
score = log₁₀(#PubMedIDs) * log₁₀(#Occurences)
The additional log-scaling is applied to filter extremly high expressed mentions and occurences. The score is additionally normalised using the top-3 hit, to create comparable, easily filterable scores, which reduce abritrary scores.